Improved accuracy of phase-system design utilizing second-order-cone-programming

This paper first formulates a non-iterative minimax method for designing a digital all-pass (AP) phase-compensation-system (PCS) by utilizing the second-order-cone-programming (SOCP) technique, and then uses an example to show that the PCS design accuracy can be improved by appropriately selecting the initial coefficients of the digital PCS. Originally, such a minimax design is a highly non-linear minimization problem. In this paper, we approximte this highly non-linear minimization problem as an SOCP problem, and then find the sub-optimal coefficients by applying an SOCP solver. This approximated SOCP design is non-iterative and can yield better design results than other designs such as linear-programming (LP) and non-iteraitve SOCP designs.